In this essay, I set out to sketch a model of "theoretical explanation with ontological commitments ". I do not claim that this model is applicable to the use of 'explanation' in daily life, nor that it is applicable to all scientific context. I only claim that it has genuine meaning in large portions of the natural sciences, particularly in physics, and more generally in those disciplines that have been more or less systematically mathematised. My overall thesis, which I shall try to articulate here, is simply the following: at least in many contexts in the theoretical sciences, explanation takes on the guise of inserting a structure of data into a theoretical model, some of whose components, precisely the theoretical ones, might have a causal interpretation.